基于视觉灵敏度及粗集的彩色图像滤波算法

为消除人眼视觉差别对彩色图像滤波的影响,提出了一种新的滤波算法VRDDF(visual sensitivity rough set directional distance filter)。对于被噪声污染的彩色图像,先用粗集理论对噪声做初步检测,在此基础上结合视觉感知灵敏度对噪声点做了进一步检测,有效的降低了误检率。再结合空间距离函数对DDF(directional distance filter)滤波算法作了合理加权改进,并作为后续滤波算法进行滤波,改善了滤波效果。该算法只对噪声点进行滤除而对非噪声点保持其值不变,可有效防止噪声传播。实验结果表明,该滤波算法和现有的滤波算法相比,在噪声滤除率、颜色保持,特别是在保护边缘细节方面均有提高,峰值信噪比平均可提高10%。

Filtering algorithm based on visual sensitivity and rough set for color image

To eliminate the influence of visual difference for color image filtering, a new filtering algorithm named visual sensitivity rough set directional distance filter （VRDDF） was proposed in this paper. The filtering algorithm used rough set theory to detect noise points for a polluted color image first, then combined with visual sensitivity to do further detection. It could reduce the false detection rate effectively. In order to improve the efficiency of the fil tering, this paper made reasonable weighted to directional distance filter （DDF） based on the space distance per ception function. Then it was used to filter noise pixels only, and the value of pixels, which were not polluted, was kept, therefore, the noise transmission could be prevented. The experiment results show that VRDDF is better than existing filtering algorithm in color keeping, noise filtering rate, especially in edge detail preservation, and the aver age peak signal to noise ratio canhe increased by 10%.